[GitHub] [tvm] ymwangg commented on a change in pull request #7441: [Frontend][Tensorflow] Add unique operator

[GitBox <gi...@apache.org>]

ymwangg commented on a change in pull request #7441:
URL: https://github.com/apache/tvm/pull/7441#discussion_r574881936



##########
File path: src/relay/op/algorithm/unique.cc
##########
@@ -0,0 +1,147 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied.  See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+
+/*!
+ * \file unique.cc
+ * \brief The unique operator
+ */
+#include <dlpack/dlpack.h>
+#include <tvm/relay/attrs/algorithm.h>
+#include <tvm/relay/op.h>
+#include <tvm/relay/op_attr_types.h>
+#include <tvm/runtime/data_type.h>
+
+namespace tvm {
+namespace relay {
+
+bool UniqueRel(const Array<Type>& types, int num_inputs, const Attrs& attrs,
+               const TypeReporter& reporter) {
+  // types: [data, result]
+  ICHECK_EQ(types.size(), 2) << "Unique: expect 2 types but " << types.size() << " provided";
+  ICHECK_EQ(num_inputs, 1) << "Unique: expect 1 inputs but " << num_inputs << " provided";
+  auto data = types[0].as<TensorTypeNode>();
+  if (data == nullptr) {
+    ICHECK(types[0].as<IncompleteTypeNode>())
+        << "Unique: expect input type to be TensorType but get " << types[0];
+    return false;
+  }
+  std::vector<Type> fields;
+  fields.push_back(TensorType(data->shape, data->dtype));

Review comment:
       The main challenge is the shape function. This operator requires data structures like set to efficiently calculate the number of unique elements. I don't think that's supported in hybrid script. Even if we can do it and get the shape function. It looks weird that both shape function and the compute function do the same computations.




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org